1. Field of the Invention
The field of the invention is data processing, or, more specifically, methods, apparatus, and products for reducing power consumption while synchronizing a plurality of compute nodes during execution of a parallel application.
2. Description of Related Art
The development of the EDVAC computer system of 1948 is often cited as the beginning of the computer era. Since that time, computer systems have evolved into extremely complicated devices. Today's computers are much more sophisticated than early systems such as the EDVAC. Computer systems typically include a combination of hardware and software components, application programs, operating systems, processors, buses, memory, input/output (‘I/O’) devices, and so on. As advances in semiconductor processing and computer architecture push the performance of the computer higher and higher, more sophisticated computer software has evolved to take advantage of the higher performance of the hardware, resulting in computer systems today that are much more powerful than just a few years ago.
Parallel computing is an area of computer technology that has experienced advances. Parallel computing is the simultaneous execution of the same task (split up and specially adapted) on multiple processors in order to obtain results faster. Parallel computing is based on the fact that the process of solving a problem usually can be divided into smaller tasks, which may be carried out simultaneously with some coordination.
Parallel computers execute applications that include both parallel algorithms and serial algorithms. A parallel algorithm can be split up to be executed a piece at a time on many different processing devices, and then put back together again at the end to get a data processing result. Some algorithms are easy to divide up into pieces. Splitting up the job of checking all of the numbers from one to a hundred thousand to see which are primes could be done, for example, by assigning a subset of the numbers to each available processor, and then putting the list of positive results back together. In this specification, the multiple processing devices that execute the algorithms of an application are referred to as ‘compute nodes.’ A parallel computer is composed of compute nodes and other processing nodes as well, including, for example, input/output (‘I/O’) nodes, and service nodes.
Parallel algorithms are valuable because it is faster to perform some kinds of large computing tasks via a parallel algorithm than it is via a serial (non-parallel) algorithm, because of the way modern processors work. It is far more difficult to construct a computer with a single fast processor than one with many slow processors with the same throughput. There are also certain theoretical limits to the potential speed of serial processors. On the other hand, every parallel algorithm has a serial part and so parallel algorithms have a saturation point. After that point adding more processors does not yield any more throughput but only increases the overhead and cost.
Parallel algorithms are designed also to optimize one more resource—the data communications requirements among the nodes of a parallel computer. There are two ways parallel processors communicate, shared memory or message passing. Shared memory processing needs additional locking for the data and imposes the overhead of additional processor and bus cycles and also serializes some portion of the algorithm.
Message passing processing uses high-speed data communications networks and message buffers, but this communication adds transfer overhead on the data communications networks as well as additional memory need for message buffers and latency in the data communications among nodes. Designs of parallel computers use specially designed data communications links so that the communication overhead will be small but it is the parallel algorithm that decides the volume of the traffic.
Many data communications network architectures are used for message passing among nodes in parallel computers. Compute nodes may be organized in a network as a ‘torus’ or ‘mesh,’ for example. Also, compute nodes may be organized in a network as a tree. A torus network connects the nodes in a three-dimensional mesh with wrap around links. Every node is connected to its six neighbors through this torus network, and each node is addressed by its x,y,z coordinate in the mesh. In such a manner, a torus network lends itself to point to point operations. In a tree network, the nodes typically are organized in a binary tree arrangement: each node has a parent, and two children (although some nodes may only have zero children or one child, depending on the hardware configuration). In computers that use a torus and a tree network, the two networks typically are implemented independently of one another, with separate routing circuits, separate physical links, and separate message buffers. A tree network provides high bandwidth and low latency for certain collective operations, message passing operations where all compute nodes participate simultaneously, such as, for example, an allgather.
As mentioned above, compute nodes connected through such data communications operate to process parallel applications. Each compute node typically processes its own set of data according to a parallel algorithm specified by the parallel application. Because each compute node processes data independently, some compute nodes may process instructions faster or slower than other compute nodes. The compute nodes that process instructions faster than other compute nodes tend to be processing data according to the parallel algorithm at a point further along in the execution sequence than the other compute nodes. The compute nodes that process instructions slower than other compute nodes tend to be processing data at a point in the execution sequence that trails the other compute nodes. Because a particular portion of a parallel algorithm often requires that all of the compute nodes begin processing the particular portion at the same time, such a parallel algorithm may specify that all of the compute nodes be synchronized before processing that particular portion of the algorithm. To synchronize a set of compute nodes, software designers often use a barrier operation. A barrier operation for a set of compute nodes prevents each compute node from processing beyond a particular point in a parallel algorithm until all of the other compute nodes reach the same point in the algorithm. A barrier operation may be implemented using, for example, the MPI_BARRIER function described in the Message Passing Interface (‘MPI’) specification that is promulgated by the MPI Forum. Because a barrier operation forces most of the compute nodes in a particular set to wait for the slowest compute nodes before processing of the parallel algorithm may continue, significant amounts of power are often wasted by compute nodes that idle while waiting for the slowest compute nodes to complete the barrier operation.
In addition to a barrier operation, there are other circumstances in which one or more compute nodes idly wait for processing to complete on another compute node before continuing data processing along an application's execution sequence. When an application executing on a plurality of compute nodes instructs one of the compute nodes to perform I/O operations, often the remaining nodes do not continue processing the application until the singular node completes performance of the I/O operations. As with the barrier operation, the remaining nodes waste significant amounts of power while waiting for the singular node to complete performance of the I/O operations.